2. 8- 2
P: the class of problems which can be solved
by a deterministic polynomial algorithm.
NP : the class of decision problem which can
be solved by a non-deterministic polynomial
algorithm.
NP-hard: the class of problems to which every
NP problem reduces.
NP-complete (NPC): the class of problems
which are NP-hard and belong to NP.
3. 8- 3
Some concepts of NPC
Definition of reduction: Problem A reduces to
problem B (A ∝ B) iff A can be solved by a
deterministic polynomial time algorithm using
a deterministic algorithm that solves B in
polynomial time.
Up to now, none of the NPC problems can be
solved by a deterministic polynomial time
algorithm in the worst case.
It does not seem to have any polynomial time
algorithm to solve the NPC problems.
4. 8- 4
The theory of NP-completeness always
considers the worst case.
The lower bound of any NPC problem seems
to be in the order of an exponential function.
Not all NP problems are difficult. (e.g. the
MST problem is an NP problem.)
If A, B ∈ NPC, then A ∝ B and B ∝ A.
Theory of NP-completeness:
If any NPC problem can be solved in polynomial
time, then all NP problems can be solved in
polynomial time. (NP = P)
5. 8- 5
Decision problems
The solution is simply “Yes” or “No”.
Optimization problems are more difficult.
e.g. the traveling salesperson problem
Optimization version:
Find the shortest tour
Decision version:
Is there a tour whose total length is less than or
equal to a constant c ?
6. 8- 6
Solving an optimization problem by a
decision algorithm :
Solving TSP optimization
problem by a decision
algorithm :
Give c1
and test (decision algorithm)
Give c2
and test (decision algorithm)
Give cn
and test (decision algorithm)
We can easily find the smallest ci
7. 8- 7
The satisfiability problem
The satisfiability problem
The logical formula :
x1
v x2
v x3
& - x1
& - x2
the assignment :
x1
← F , x2
← F , x3
← T
will make the above formula true .
(-x1
, -x2
, x3
) represents x1
← F , x2
← F , x3
← T
8. 8- 8
If there is at least one assignment which
satisfies a formula, then we say that this
formula is satisfiable; otherwise, it is
unsatisfiable.
An unsatisfiable formula :
x1
v x2
& x1
v -x2
& -x1
v x2
& -x1
v -x2
9. 8- 9
Definition of the satisfiability problem: Given
a Boolean formula, determine whether this
formula is satisfiable or not.
A literal : xi
or -xi
A clause : x1
v x2
v -x3
≡ Ci
A formula : conjunctive normal form (CNF)
C1
& C2
& … & Cm
10. 8- 10
Resolution principle
C1
: x1
v x2
C2
: -x1
v x3
⇒ C3
: x2
v x3
From C1
&C2
, we can
obtain C3
, and C3
can
be added into the
formula.
The formula becomes:
C1
&C2
&C3
The resolution principle
x1
x2
x3 C1
&C2
C3
0 0 0 0 0
0 0 1 0 1
0 1 0 1 1
0 1 1 1 1
1 0 0 0 0
1 0 1 1 1
1 1 0 0 1
1 1 1 1 1
11. 8- 11
Another example of resolution principle
C1
: -x1
v -x2
v x3
C2
: x1
v x4
⇒ C3
: -x2
v x3
v x4
If no new clauses can be deduced, then
it is satisfiable.
-x1
v -x2
v x3
(1)
x1
(2)
x2
(3)
(1) & (2) -x2
v x3
(4)
(4) & (3) x3
(5)
(1) & (3) -x1
v x3
(6)
12. 8- 12
If an empty clause is deduced, then it is
unsatisfiable.
- x1
v -x2
v x3
(1)
x1
v -x2
(2)
x2
(3)
- x3
(4)
⇓ deduce
(1) & (2) -x2
v x3
(5)
(4) & (5) -x2
(6)
(6) & (3) □ (7)
13. 8- 13
Semantic tree
In a semantic tree, each
path from the root to a
leaf node represents a
class of assignments.
If each leaf node is
attached with a clause,
then it is unsatisfiable.
14. 8- 14
Nondeterministic algorithms
A nondeterminstic algorithm consists of
phase 1: guessing
phase 2: checking
If the checking stage of a nondeterministic
algorithm is of polynomial time-complexity, then
this algorithm is called an NP (nondeterministic
polynomial) algorithm.
NP problems : (must be decision problems)
e.g. searching, MST
sorting
satisfiability problem (SAT)
traveling salesperson problem (TSP)
15. 8- 15
Decision problems
Decision version of sorting:
Given a1
, a2
,…, an
and c, is there a permutation
of ai
′
s ( a1
′
, a2
′
, … ,an
′
) such that∣a2
′
–a1
′
∣+∣a3
′
–a2
′
∣+
… +∣an
′
–an-1
′
∣ < c ?
Not all decision problems are NP problems
E.g. halting problem :
Given a program with a certain input data, will
the program terminate or not?
NP-hard
Undecidable
16. 8- 16
Nondeterministic operations
and functions
[Horowitz 1998]
Choice(S) : arbitrarily chooses one of the elements in set
S
Failure : an unsuccessful completion
Success : a successful completion
Nonderministic searching algorithm:
j ← choice(1 : n) /* guessing */
if A(j) = x then success /* checking */
else failure
17. 8- 17
A nondeterministic algorithm terminates
unsuccessfully iff there does not exist a set of
choices leading to a success signal.
The time required for choice(1 : n) is O(1).
A deterministic interpretation of a non-
deterministic algorithm can be made by
allowing unbounded parallelism in
computation.
18. 8- 18
Nondeterministic sorting
B ← 0
/* guessing */
for i = 1 to n do
j ← choice(1 : n)
if B[j] ≠ 0 then failure
B[j] = A[i]
/* checking */
for i = 1 to n-1 do
if B[i] > B[i+1] then failure
success
19. 8- 19
Nondeterministic SAT
/* guessing */
for i = 1 to n do
xi ← choice( true, false )
/* checking */
if E(x1, x2, … ,xn) is true then success
else failure
20. 8- 20
Cook’s theorem
NP = P iff the satisfiability
problem is a P problem.
SAT is NP-complete.
It is the first NP-complete
problem.
Every NP problem reduces
to SAT.
Stephen Arthur
Cook
(1939~)
21. 8- 21
Transforming searching to SAT
Does there exist a number in { x(1),
x(2), …, x(n) }, which is equal to 7?
Assume n = 2.
nondeterministic algorithm:
i = choice(1,2)
if x(i)=7 then SUCCESS
else FAILURE
23. 8- 23
CNF (conjunctive normal form) :
i=1 v i=2 (1)
i≠1 v i≠2 (2)
x(1)≠7 v i≠1 v SUCCESS (3)
x(2)≠7 v i≠2 v SUCCESS (4)
x(1)=7 v i≠1 v FAILURE (5)
x(2)=7 v i≠2 v FAILURE (6)
-FAILURE v -SUCCESS (7)
SUCCESS (8)
x(1)=7 (9)
x(2)≠7 (10)
24. 8- 24
Satisfiable at the following assignment :
i=1 satisfying (1)
i≠2 satisfying (2), (4) and (6)
SUCCESS satisfying (3), (4) and (8)
−FAILURE satisfying (7)
x(1)=7 satisfying (5) and (9)
x(2)≠7 satisfying (4) and (10)
25. 8- 25
The semantic tree
i=1 v i=2 (1)
i≠1 v i≠2 (2)
x(1)≠7 v i≠1 v SUCCESS (3)
x(2)≠7 v i≠2 v SUCCESS (4)
x(1)=7 v i≠1 v FAILURE (5)
x(2)=7 v i≠2 v FAILURE (6)
-FAILURE v -SUCCESS (7)
SUCCESS (8)
x(1)=7 (9)
x(2)≠7 (10)
26. 8- 26
Searching for 7, but x(1)≠7, x(2)≠7
CNF (conjunctive normal form) :
i=1 v i=2 (1)
i≠1 v i≠2 (2)
x(1)≠7 v i≠1 v SUCCESS (3)
x(2)≠7 v i≠2 v SUCCESS (4)
x(1)=7 v i≠1 v FAILURE (5)
x(2)=7 v i≠2 v FAILURE (6)
SUCCESS (7)
-SUCCESS v -FAILURE (8)
x(1) ≠ 7 (9)
x(2) ≠ 7 (10)
27. 8- 27
Apply resolution principle :
(9) & (5) i≠1 v FAILURE (11)
(10) & (6) i≠2 v FAILURE (12)
(7) & (8) -FAILURE (13)
(13) & (11) i≠1 (14)
(13) & (12) i≠2 (15)
(14) & (1) i=2 (11)
(15) & (16) □ (17)
We get an empty clause ⇒ unsatisfiable
⇒ 7 does not exit in x(1) or x(2).
28. 8- 28
CNF:
i=1 v i=2 (1)
i≠1 v i≠2 (2)
x(1)≠7 v i≠1 v S U C C E S S (3 )
x(2)≠7 v i≠2 v S U C C E S S (4 )
x(1)=7 v i≠1 v FA I L U R E ( 5 )
x(2)=7 v i≠2 v FA I L U R E ( 6 )
SUCCESS (7)
-SUCCESS v -FAILURE (8)
x(1)=7 (9)
x(2)=7 (10)
Searching for 7, where x(1)=7, x(2)=7
30. 8- 30
The node cover problem
Def: Given a graph G=(V, E), S is the node
cover if S ⊆ V and for every edge (u, v) ∈ E,
either u ∈ S or v ∈ S.
node cover :
{1, 3}
{5, 2, 4}
Decision problem : ∃ S ∋ | S | ≤ K ?
31. 8- 31
Transforming the node cover
problem to SAT
BEGIN
i1 ← choice({1, 2, …, n})
i2 ← choice({1, 2, …, n} – {i1})
ik ← choice({1, 2, …, n} – {i1, i2, …, ik-1}).
For j=1 to m do
BEGIN
if ej is not incident to one of νit
(1≤t≤k)
then FAILURE
END
SUCCESS
32. 8- 32
i1 = 1 v i1 = 2… v i1 = n
(i1≠ 1→i1=2 v i1=3…v i1 = n)
i2 = 1 v i2 = 2… v i2 = n
ik = 1 v ik = 2… v ik = n
i1 ≠1 v i2 ≠1 (i1=1 →i2≠1 & ... & ik≠1)
i1 ≠1 v i3 ≠1
ik-1 ≠n v ik ≠n
νi1
∈e1 v νi2
∈e1 v … v νik
∈e1 v FAILURE
(νi1
∉e1&νi2
∉e1&…&νik
∉e1→ Failure)
νi1
∈e2 v νi2
∈e2 v … v νik
∈e2 v FAILURE
νi1
∈em v νi2
∈em v … v νik
∈em v FAILURE
SUCCESS
CN
F:
(To be continued)
34. 8- 34
SAT is NP-complete
(1) SAT has an NP algorithm.
(2) SAT is NP-hard:
Every NP algorithm for problem A can be
transformed in polynomial time to SAT
[Horowitz 1998] such that SAT is satisfiable
if and only if the answer for A is “YES”.
That is, every NP problem ∝ SAT .
By (1) and (2), SAT is NP-complete.
35. 8- 35
Proof of NP-Completeness
To show that A is NP-complete
(I) Prove that A is an NP problem.
(II) Prove that ∃ B ∈ NPC, B ∝ A.
⇒ A ∈ NPC
Why ?
36. 8- 36
3-satisfiability problem (3-
SAT)
Def: Each clause contains exactly three
literals.
(I) 3-SAT is an NP problem (obviously)
(II) SAT ∝ 3-SAT
Proof:
(1) One literal L1
in a clause in SAT :
in 3-SAT :
L1
v y1
v y2
L1
v -y1
v y2
L1
v y1
v -y2
L1
v -y1
v -y2
37. 8- 37
(2) Two literals L1
, L2
in a clause in SAT :
in 3-SAT :
L1
v L2
v y1
L1
v L2
v -y1
(3) Three literals in a clause : remain unchanged.
(4) More than 3 literals L1
, L2
, …, Lk
in a clause :
in 3-SAT :
L1
v L2
v y1
L3
v -y1
v y2
Lk-2
v -yk-4
v yk-3
Lk-1
v Lk
v -yk-3
38. 8- 38
SAT transform
3-SAT
S S′
The instance S′ in 3-
SAT :
x1
v x2
v y1
x1
v x2
v -y1
-x3
v y2
v y3
-x3
v -y2
v y3
-x3
v y2
v -y3
-x3
v -y2
v -y3
x1
v -x2
v y4
x3
v -y4
v y5
-x4
v -y5
v y6
x5
v x6
v -y6
An instance S in SAT :
x1
v x2
-x3
x1
v -x2
v x3
v -x4
v x5
v x6
Example of transforming SAT to 3-
SAT
39. 8- 39
Proof : S is satisfiable ⇔ S′ is satisfiable
“⇒”
≤ 3 literals in S (trivial)
consider ≥ 4 literals
S : L1
v L2
v … v Lk
S′: L1
v L2
v y1
L3
v -y1
v y2
L4
v -y2
v y3
Lk-2
v -yk-4
v yk-3
Lk-1
v Lk
v -yk-3
40. 8- 40
S is satisfiable ⇒ at least Li
= T
Assume : Lj
= F ∀ j ≠ i
assign : yi-1
= F
yj
= T ∀ j < i-1
yj
= F ∀ j > i-1
( Li
v -yi-2
v yi-1
)
⇒ S′ is satisfiable.
“⇐”
If S′ is satisfiable, then assignment satisfying
S′ can not contain yi
’s only.
⇒ at least one Li
must be true.
(We can also apply the resolution principle).
Thus, 3-SAT is NP-complete.
41. 8- 41
Comment for 3-SAT
If a problem is NP-complete, its special cases
may or may not be NP-complete.
42. 8- 42
Chromatic number decision
problem (CN)
Def: A coloring of a graph G=(V, E) is a function
f : V → { 1, 2, 3,…, k } such that if (u, v) ∈ E, then
f(u)≠f(v). The CN problem is to determine if G has a
coloring for k.
<Theorem> Satisfiability with at most 3 literals per
clause (SATY) ∝ CN.
3-colorable
f(a)=1, f(b)=2, f(c)=1
f(d)=2, f(e)=3
43. 8- 43
Proof :
instance of SATY :
variable : x1, x2, …, xn , n ≥ 4
clause : c1, c2, …, cr
instance of CN :
G=(V, E)
V={ x1, x2, …, xn }∪{ -x1, -x2, …, -xn }
∪{ y1, y2, …, yn }∪{ c1, c2, …, cr }
newly added
E={ (xi, -xi) | 1≤ i ≤ n }∪{ (yi, yj) | i ≠ j }
∪{ (yi, xj) | i ≠ j }∪{ (yi, -xj) | i ≠ j }
∪{ (xi, cj) | xi ∉ cj }∪{ (-xi, cj) | -xi ∉ cj }
SATY ∝ CN
44. 8- 44
x1
v x2
v x3
(1)
-x3
v -x4
v x2
(2)
⇓
Example of SATY ∝ CN
True assignment:
x1
=T
x2
=F
x3
=F
x4
=T
E={ (xi
, -xi
) | 1≤ i ≤ n }∪{ (yi
, yj
) | i ≠ j }
∪{ (yi
, xj
) | i ≠ j }∪{ (yi
, -xj
) | i ≠ j }
∪{ (xi
, cj
) | xi
∉ cj
}∪{ (-xi
, cj
) | -xi
∉ cj
}
45. 8- 45
Satisfiable ⇔ n+1 colorable
“⇒”
(1) f(yi
) = i
(2) if xi
= T, then f(xi
) = i, f(-xi
) = n+1
else f(xi
) = n+1, f(-xi
) = i
(3)if xi
in cj
and xi
= T, then f(cj
) = f(xi
)
if -xi
in cj
and -xi
= T, then f(cj
) = f(-xi
)
( at least one such xi
)
Proof of SATY ∝ CN
46. 8- 46
“⇐”
(1) yi
must be assigned with color i.
(2) f(xi
) ≠ f(-xi
)
either f(xi
) = i and f(-xi
) = n+1
or f(xi
) = n+1 and f(-xi
) = i
(3) at most 3 literals in cj
and n ≥ 4
⇒ at least one xi
, ∋ xi
and -xi
are not in cj
⇒ f(cj
) ≠ n+1
(4) if f(cj
) = i = f(xi
), assign xi
to T
if f(cj
) = i = f(-xi
), assign -xi
to T
(5) if f(cj
) = i = f(xi
) ⇒ (cj
, xi
) ∉ E
⇒ xi
in cj
⇒ cj
is true
if f(cj
) = i = f(-xi
) ⇒ similarly
47. 8- 47
Set cover decision problem
Def: F = {Si
} = { S1
, S2
, …, Sk
}
= { u1
, u2
, …, un
}
T is a set cover of F if T ⊆ F and
The set cover decision problem is to determine if F has
a cover T containing no more than c sets.
Example
F = {(u1
, u3
), (u2
, u4
), (u2
, u3
), (u4
), (u1
, u3
, u4
)}
s1
s2
s3
s4
s5
T = { s1
, s3
, s4
} set cover
T = { s1
, s2
} set cover, exact cover
FS
i
i
S
∈
FS
i
TS
i
ii
SS
∈∈
=
48. 8- 48
Exact cover problem
(Notations same as those in set cover.)
Def: To determine if F has an exact cover T,
which is a cover of F and the sets in T are
pairwise disjoint.
<Theorem> CN ∝ exact cover
(No proof here.)
49. 8- 49
Sum of subsets problem
Def: A set of positive numbers A = { a1
, a2
, …,
an
}
a constant C
Determine if ∃ A′ ⊆ A ∋
e.g. A = { 7, 5, 19, 1, 12, 8, 14 }
C = 21, A′ = { 7, 14 }
C = 11, no solution
<Theorem> Exact cover ∝ sum of subsets.
Ca
Aa
i
i
=∑′∈
50. 8- 50
Proof :
instance of exact cover :
F = { S1
, S2
, …, Sk
}
instance of sum of subsets :
A = { a1
, a2
, …, ak
} where
where eij
= 1 if uj
∈ Si
eij
= 0 if otherwise.
Why k+1? (See the example on the next page.)
Exact cover ∝ sum of subsets
{ }
FS
ni
i
uuuS
∈
= ...,,2,1
j
nj
iji kea )1(
1
+= ∑≤≤
kkkkC n
nj
j
/)1)1)((1()1(
1
−++=+= ∑≤≤
52. 8- 52
Partition problem
Def: Given a set of positive numbers A =
{ a1
,a2
,…,an
},
determine if ∃ a partition P, ∋
e. g. A = {3, 6, 1, 9, 4, 11}
partition : {3, 1, 9, 4} and {6, 11}
<Theorem> sum of subsets ∝ partition
∑∑ ∉∈
=
Pa
i
Pa
i
ii
aa
53. 8- 53
Sum of subsets ∝ partition
proof :
instance of sum of subsets :
A = { a1, a2, …, an }, C
instance of partition :
B = { b1, b2, …, bn+2 }, where bi = ai, 1≤ i ≤ n
bn+1 = C+1
bn+2 = ( ∑ ai )+1−C
1≤i≤n
C = ∑ai ⇔ ( ∑ai )+bn+2 = ( ∑ai )+bn+1
ai∈S ai∈S ai∉S
⇔ partition : { bi | ai∈S}∪{bn+2}
and { bi | ai∉S }∪{bn+1}
S
S’
A C
54. 8- 54
Why bn+1
= C+1 ? why not bn+1
= C ?
To avoid bn+1
and bn+2
to be partitioned into
the same subset.
55. 8- 55
Bin packing problem
Def: n items, each of size ci
, ci
> 0
Each bin capacity : C
Determine if we can assign the items into
k bins, ∋ ∑ci
≤ C , 1≤j≤k.
i∈binj
<Theorem> partition ∝ bin packing.
56. 8- 56
VLSI discrete layout problem
Given: n rectangles, each with height hi
(integer)
width wi
and an area A
Determine if there is a placement of the n
rectangles within the area A according to the rules :
1. Boundaries of rectangles parallel to x axis or y
axis.
2. Corners of rectangles lie on integer points.
3. No two rectangles overlap.
4. Two rectangles are separated by at least a unit
distance.
(See the figure on the next page.)
57. 8- 57
A Successful Placement
<Theorem> bin packing ∝ VLSI discrete layout.
58. 8- 58
Max clique problem
Def: A maximal complete subgraph of a graph
G=(V,E) is a clique. The max (maximum) clique
problem is to determine the size of a largest
clique in G.
e. g.
<Theorem> SAT ∝ clique decision problem.
maximal cliques :
{a, b}, {a, c, d}
{c, d, e, f}
maximum clique :
(largest)
{c, d, e, f}
59. 8- 59
Node cover decision problem
Def: A set S ⊆ V is a node cover for a graph
G = (V, E) iff all edges in E are incident to at
least one vertex in S. ∃ S, ∋ |S| ≤ K ?
<Theorem> clique decision problem ∝
node cover decision problem.
(See proof on the next page.)
60. 8- 60
Clique decision ∝ node cover
decision
G=(V,E) : clique Q of size k (Q⊆V)
G’=(V,E’) : node cover S of size n-k, S=V-Q
where E’={(u,v)|u∈V, v ∈V and (u,v)∉E}
61. 8- 61
Hamiltonian cycle problem
Def: A Hamiltonian cycle is a round trip path
along n edges of G which visits every vertex
once and returns to its starting vertex.
e.g.
Hamiltonian cycle : 1, 2, 8, 7, 6, 5, 4, 3, 1.
<Theorem> SAT ∝ directed Hamiltonian cycle
( in a directed graph )
62. 8- 62
Traveling salesperson problem
Def: A tour of a directed graph G=(V,
E) is a directed cycle that includes every
vertex in V. The problem is to find a tour
of minimum cost.
<Theorem> Directed Hamiltonian cycle ∝
traveling salesperson decision problem.
(See proof on the next page.)
64. 8- 64
0/1 knapsack problem
Def: n objects, each with a weight wi
> 0
a profit pi
> 0
capacity of knapsack : M
Maximize ∑pi
xi
1≤i≤n
Subject to ∑wi
xi
≤ M
1≤i≤n
xi
= 0 or 1, 1≤ i ≤n
Decision version :
Given K, ∃ ∑pi
xi
≥ K ?
1≤i≤n
Knapsack problem : 0 ≤ xi
≤ 1, 1≤ i ≤n.
<Theorem> partition ∝ 0/1 knapsack decision
problem.
65. 8- 65
Refer to Sec. 11.3, Sec. 11.4 and its
exercises of [Horowitz 1998] for the proofs of
more NP-complete problems.
[[Horowitz 1998] E. Howowitz, S. Sahni and S.
Rajasekaran, Computer Algorithms, Computer
Science Press, New York, 1998, 「台北圖書」代
理 , 02-23625376